Intuitive priors generator

This tool helps you define a beta distribution by using an intuitive description of uncertainty, rather than unintuitive parameters.

Use this tool when:

• You are studying a proportion or percentage, such as task completion rates, or rate of agreement with a survey item.
• You have an intuitive sense of the likely value of this proportion: a best guess, and a general range of what reasonable outcomes could be.
• For example, you might guess that 80% of people will complete a given task, and you'd be surprised if fewer than 70% or more than 90% will complete.
• You can use this intuition sense to define a beta distribution that can be used as a prior in a Bayesian analysis.

Why use a prior distribution:

• By ruling out unreasonable values, you can sharpen your statistical inferences. For example, you might be able to report 95% chance of being directionally correct rather than 90% chance, because you reduced some uncertainty.
• Using priors is mathematically equivalent to collecting more data, so you can think of this as increasing the sample size of your study, for free.
• If you are uncomfortable with the idea, you can use a weak prior: Just provide a very large range of reasonable values.